Huffman encoding is a popular variable-length algorithm based on the fact that certain alphabets are seen more frequently than others. The idea is simple, but I've seldom seen it being used while writing user stories.
Not all features of a product are going to be equally popular. But we still build with the assumption that every feature is going to be used by everyone.
I've never seen the anticipated frequency of use of a feature with respect to other features documented. Every feature is thus made "equally accessible." (or why would we put it in the product in the first place?)
A healthy dose of Huffman encoding applied to user stories helps make the common things quick, the uncommon things slightly slightly less accessible, and the experimental stuff requiring many clicks. And it's not just the UX, but the product prioritisation and engineering aspects too.
In short, the infrequently used features should not be given disproportionate attention.